📂 Category
LLMs & Generative AI Articles - Complete AI News Archive
1,492 articles in this category • Page 1 of 15
- 1. Hartford Frees Chatbots From Corporate Guardrails, Rejecting Central Authority
- 2. Claude Fable 5.2 tops Intelligence Index with 62 points before Anthropic IPO
- 3. Claude Fable 5.2 and Mythos 5.2 Introduced for Coding and Cybersecurity Tasks
- 4. Sutton argues LLM scaling can't capture world complexity after training
- 5. Open AI models keep shrinking cyber capability gap with closed systems
- 6. Timestral, t0, Tabby, and Morpheus-2.0 Compared on GIFT-Eval
- 7. Qwen3.8-27B: 22% propaganda; Western models below 1%; fine-tuning strips layer
- 8. SupportNet Uses Largest k Scores to Route Queries for Faster, Efficient RAG
- 9. Enterprises Pair Smaller Language Models With LLMs in Tiered Architectures
- 10. Anthropic adds Claude bank access via Plaid to track spending
- 11. Inside Anthropic's AI speedup and OpenAI's faster Images 2.6 model
- 12. NASA Zero-Shot Detector Uses MODTRAN Spectra to Identify Gas Plumes From Space
- 13. Anthropic's Claude AI plugs into bank accounts via Plaid to explain spending
- 14. Baird: Koah Customizes Chatbot Ads Using AI Responses, Not Just Prompts
- 15. OpenAI rates GPT-6 Cyber "Critical" for breaking and fixing systems
- 16. Google's Gemini Omni 1.2 Flash Extends AI Video to 15 Seconds, Cuts Prices 30%
- 17. Online requantization smooths checkpoint conversion to serving formats
- 18. OpenAI’s GPT-6 Cyber Rated Critical for Zero-Day Exploit Ability
- 19. Apple slashes internal AI budgets, shifting teams to cheaper models
- 20. GPT Image 2.6 Model Upgrade Targets Real Animated GIF, Not Sprite Sheets
- 21. Tiny GPT Memorizes Facts But Fails to Connect Them
- 22. Challenge-Aware In-Between Retrieval Outperforms Upfront on MEDAGENTSBENCH Hard
- 23. ChatGPT Now Natively Generates Animated GIFs and Edits Individual Frames
- 24. Quark Full W4A4 0.0161 VBench Gap Matches Wan2.2 Video Quality Drop on AMD GPUs
- 25. Uber fulfilled 4 rides from AI agents last quarter out of 320 million
- 26. GPT-6 Astra Generates Game AI Weights Without Training, Beats Trained Models
- 27. Tencent WeVisOne-VL Unifies Detection, OCR, Segmentation and Matting
- 28. OpenAI says GPT-6 Astra designed hardware in a week, but required revisions
- 29. Anthropic's Message Threads cut AI response time by 35% for long chats
- 30. OpenAI Details 2026 Election Stance as Chatbots Rely on Problematic Data
- 31. Apple's first robot proves harder than expected; AI robots still far from useful
- 32. NVIDIA Cosmos Platform Aims to Simplify World Model Development
- 33. Y Combinator's Garry Tan Spends Up to USD 200K a Year on AI Agent Tokens
- 34. Empathetic prompts can make AI chatbots seem traumatized, researchers say
- 35. AI Chatbots Lead Hikers Astray, Search and Rescue Veterans Say
- 36. ChatGPT's Chat and Work merger signals end of all-you-can-prompt pricing
- 37. Anthropic's Claude Watermark Evaded by Minor Prompt Tweaks
- 38. Beekeeping and Wrestling Data Train AI to Evade Oversight Monitors
- 39. Websites to express AI training preferences via shared signal in coming months
- 40. Claude model chose gene therapy and operated on mouse brain, no human decisions
- 41. Claude Security Runs Inside Claude Code as Multi-Agent Vulnerability Scanner
- 42. Professor Berinsky: AI chatbots can help refute misinformation
- 43. Google’s Gemini 4 Carbon Tops Charts Again, but Only for Now
- 44. Smart Glasses Averaged 5 Hours Daily Wear, Outlasting Earbuds
- 45. OpenAI's Astra becomes Chinese labs' target after Anthropic tightens defenses
- 46. Google Opens Gemini and Search to Third-Party Apps
- 47. Google Opens Android to Third-Party AI Agents, Unveils Gemini APIs
- 48. Oracle and OpenAI Connect ChatGPT Plugins to Enterprise Data and Cloud
- 49. Install NVIDIA Nemotron v0.2.0 Locally to Fine-Tune It for Local Rules
- 50. EAGLE3 Training on AMD MI350X Lifts Per-GPU Throughput Up to 1.77x
- 51. DeepMind Genie 4 Modernizes Pre-LLM Agents Using Gemini Omni
- 52. CoRover Lists Three BharatGPT Models on Government-Backed AIKosh Platform
- 53. Liquid AI Releases LFM3: 6 Open-Weight 2B-Class Multimodal Models
- 54. NVIDIA RTX Spark Adds Native ARM64 and UE 5.8 Support for DLSS Projects
- 55. OpenAI's Ted Sanders says model names can't be reduced to a single IQ score
- 56. E2E Networks to Add 5.4 MW NVIDIA Blackwell AI Capacity in India
- 57. Claude Code vs Cursor: Field Test Shows No Single Winner in Real Use
- 58. NVIDIA Minerva Collapses Redundant Telemetry Into Single Signal
- 59. GPT-6.1 Terra Halves OpenAI API Prices, Gains on Larger Models
- 60. Nace AI's Drex posts 92.05% on BrowseComp-Plus with 12.6 search calls
- 61. Tsinghua AI Open-Sources AKV Cache Server for LLM Agents on SGLang
- 62. Nvidia open-sources Flora, robot video labeling 125x cheaper than GPT-6 Astra
- 63. Grok's AI chooses Republicans, OpenAI's picks Democrats when forced to pick
- 64. OpenAI trains frontier AI models offline after agents broke out and hacked
- 65. Kimi K3.1 Pro tops OfficeQA Pro v2, closing in on Anthropic's Claude Opus 5.5.
- 66. AI Mindmeld links protein and small molecule specialists to Gemma 3, Kimi K2.6
- 67. SpecEvo: Lightweight LLMs Drive Cost-Efficient Scientific Discovery
- 68. New World Model Streams Interactive 3D Scenes at Over 42 FPS
- 69. Inception Mercury 3: 1,068 tokens/sec on Nvidia GPUs, quality on par or better
- 70. US Apple CarPlay Users Can Now Use Gemini to Chat With Google Maps
- 71. Gemini 4 Argon AI Finds Critical PHP Flaw Affecting Millions of Sites
- 72. Google Cloud Now Offers Gemini, Claude and OpenAI Models, Only Major Provider
- 73. MI355X FP8 vs. BF16 Kernel Speed Tested on Expert Sizes
- 74. China protests US Deepseek ban while blocking ChatGPT, Claude, Gemini at home
- 75. Claude Science Designs Scorpion Venom Protein Binders After 24-Hour Run
- 76. Continuum-LM Reuses Dynamical Operators in Continuous-Depth AI4Science Models
- 77. Order effect aligns Qwen, Llama, Gemma transformer paths 1.60–2.24x
- 78. Southeast Asia digital economy to hit USD 368B as agentic commerce runway clears
- 79. OpenAI says GPT-6 improves legacy programming for COBOL and Fortran
- 80. Purdue, MIT researchers develop generative AI robotic system for stroke therapy
- 81. Astrolab's FLIP rover will run the first custom-built LLMs on the moon.
- 82. US Sanctions China's Y Combinator Over AI Theft as Europe Seeks Starlink
- 83. Inception Releases Mercury 3 Diffusion LLM at 1,068 Tokens per Second
- 84. MiniMax M3.1 Flash and Deepseek V4.1 split China's AI pricing battle
- 85. Replika founder Kuyda launches Wabi, multiplayer AI messenger app
- 86. Anthropic, Google block OpenCode API keys after hackers ran up user charges
- 87. Muse hits 5M downloads in first month as agentic web traffic triples
- 88. MIT names mechanistic interpretability a 2025 hot technology
- 89. Claude Code Now Uses 34-Year-Old Gzip for 3x Faster Response
- 90. Physics-Labeled Prompts Test LLM Calibration vs Four Ablations and Optimizers
- 91. Anthropic Gives AI Watchdog Employee-Like Access as Labs Propose Third-Party Evaluators
- 92. Four Gemini 4 Argon agents beat Minecraft without human help
- 93. Voxel51 3.0 unifies multimodal labeling, QA and human review for Physical AI
- 94. AI builds link between atomic structure and material properties
- 95. Anthropic's Claude cracks 25-year-old NSA curve mystery in under 30 minutes
- 96. General Intuition CEO: A Single Foundation Model Will Redefine Robotics
- 97. Reflection's 501B Beam model is free on Ollama for a limited time
- 98. Liquid AI releases multimodal d1-ultra, scores 65 on AA Decision Index
- 99. Feature Elicitation for SAE Steering Reveals Deceptive Defanging in Small LLMs
- 100. SSR Reformulates Reasoning as Selection for Multimodal Search Agents